Blog
Notes on applied AI
Practical insight on building and governing AI systems, for leaders and founders.New models, benchmarks and the methods we use ourselves.
11 Aug 2026 · RS Management
How to Check Whether AI Recommends Your Company
How to measure a company's visibility in ChatGPT, Claude, Gemini and other assistants: sample questions, common measurement errors and what changes the result.
Read the note →7 Aug 2026 · RS Management
How to Assess an AI Project on GitHub
The most popular AI projects on GitHub in August 2026, what star counts say about a project's maturity, what they don't and six risk signals to check in fifteen minutes.
Read the note →4 Aug 2026 · RS Management
The AI Act After 2 August 2026
Regulation 2024/1689 after the July 2026 amendment: what began to apply on 2 August, what moved to December 2027 and what it means for HR.
Read the note →21 Jul 2026 · RS Management
How to Cut the Cost of Tokens and Context
Why the AI bill grows faster than the number of requests and how to lower it: prompt caching, history summarization, RAG, output limits, routing and batch processing.
Read the note →7 Jul 2026 · RS Management
How to Connect an AI Agent to Company Systems
Comparing the MCP protocol and command-line integration: when standardization pays off, when a script is enough, what context costs and how to secure calls.
Read the note →23 Jun 2026 · RS Management
How to Write Better Prompts for AI
Five parts of a good prompt for an AI model, three before-and-after pairs from office work, a plan before execution, iterating on errors and checking before sending.
Read the note →9 Jun 2026 · RS Management
What Hardware You Need to Run AI Models Locally
Four hardware tiers for running AI models inside your own company, the software, memory requirements and the cost comparison with an API at 2026 prices.
Read the note →26 May 2026 · RS Management
How to Read AI Model Benchmarks
Why AI model rankings mislead: data contamination, errors in test sets and result variance. Four questions before quoting a score and building your own test set.
Read the note →12 May 2026 · RS Management
How to Choose Between a Commercial and an Open AI Model
When a commercial model via API wins and when an open-weights model run in-house does: data confidentiality, cost at scale, non-English quality, version control and licenses.
Read the note →28 Apr 2026 · RS Management
How to Take an AI Prototype to Production
A half-page prototype card, production data, a business owner, security from day one and a decision date: what separates a prototype that goes into use.
Read the note →14 Apr 2026 · RS Management
How to Deploy a Chatbot and Voicebot in a Large Company
What to settle before choosing a platform: knowledge base, permissions, no-answer paths, escalation, measurement and content upkeep, with the voice channel last.
Read the note →10 Mar 2026 · RS Management
What the AI Act Requires of a Company That Uses AI
A guide to the EU AI Act for leaders in regulated industries: the timeline of obligations, deployer accountability, risk categories and first steps.
Read the note →26 Feb 2026 · RS Management
When One AI Agent Stops Being Enough
Why a do-everything AI agent loses quality as its scope grows and how a team of specialized agents with an orchestrator restores control and accountability.
Read the note →25 Feb 2026 · RS Management
How to Match an AI Model to the Task
Which AI model should handle which task: three task tiers, differences in price and speed, model cascades and the threshold below which routing does not pay off.
Read the note →Not sure where to start?
A 15-minute call is enough to work out which package to start with.
Book an intro call